    ----====== HUGE_FLOAT ver 22-October-2000 ======---- 
       author: Joerg Arndt, email: arndt@jjj.de 
     compiler used: GNU C 2.95.2 19991024 (release)
        compilation date: Oct 26 2000, 20:06:47
       hfloat is online at http://www.jjj.de/hfloat/   
     ----========================================----  

USAGE:  '../bin/pi LDN WHAT RADIX FNAME'   (or  '../bin/pi --help'  for this help)
 First arg (LDN) is the (base 2) logarithm of the desired precision in LIMBs
   A LIMB is the 'superdigit' used by hfloat internally
   (the decimal precision is prec*log_10(RADIX))
   e.g. LDN=11 means 2^11=2048 LIMBS;  default is 10 (or 1024 LIMBs)
 Second arg (WHAT) chooses what to do (see below)
 Third arg (RADIX) gives radix (default is 10000)
   (usually you'll use 10000 for decimal digits, 65536 for hex digits)
 Fourth arg (FNAME) gives filename where to save the result.

 Second arg chooses what to do:
  Computations marked with '!' are for modest precisions only.
    0 => pi, quartic iteration with r=4 (default)
    1 => pi, quartic iteration with r=16
    2 => pi, 2.order iteration
    3 => pi, agm iteration
    4 => pi, agm3 iteration, faster variant
    5 => pi, agm3 iteration, slower variant
    6 => pi, derived agm iteration
    7 => pi, quintic iteration
    8 => pi, 3.order iteration
    9 => pi, 9.order iteration
   10 => pi, cubic agm iteration
   13 => pi, agm iteration, quartic variant
   14 => pi, agm3 iteration, quartic faster variant
   15 => pi, agm3 iteration, quartic slower variant
   16 => pi, agm iteration a la Schoenhage
   5N => ! pi, N-term arctan series (N=2...7)
   60 => ! pi, as arctan(1) (via sumalt)
  100 => ! eulers e  (2.7182818...) via exp(1)
  101 => ! eulers e  (2.7182818...) via series
  111 => log(2) =0.6931471805599453...  via agm
  112 => ! log(2) =0.6931471805599453...  via sumalt
  120 => ! zeta(3) =1.202056903159594...  via sumalt
  141 => sqrt(2) =1.4142135623
  161 => golden =(sqrt(5)+1)/2 =1.61803398874989...
  163 => ! exp(pi*sqrt(163)) =262537412640768743.999...
  207 => ! i**i =exp(-pi/2) =0.2078795763507619...
  314 => pi-check (run different algorithms for pi & compare the results)
  847 => agm(1,1/sqrt(2)) =0.8472130847939790866064991234...
  915 => ! catalan=0.915965594177219015 ... via sumalt()
  999 => (your) stuff in yourstuff.cc

  e.g.: ../bin/pi 11 1 65536
    use method 1, precision= 2^11=2048 (radix 65536) LIMBs
    (i.e. precision is 2048*16 bits= 4096 bytes)
  e.g.: ../bin/pi 8 0 10000  (default action)
    use method 0, precision= 2^8=256 (radix 10000) LIMBs
    (i.e. precision is 256*4=1024 decimal digits)
  e.g.: ../bin/pi 18 16 10000 result.txt
    compute pi to 1 million decimal digits (4*2^18=1M)
    use Schoenhage's agm, save result to result.txt
